164 M. Ramond*s Instruct ions for the Application of [Se1»t. 



dicular to the axis of the column. Some observers have added 

 a microscope to discern more exactly the point of contact. This 

 appears unnecessary, as we can judge without a microscope very 

 well to the 1-lOOOth of an inch ; and the adherence of the mer- 

 cury to the sides of the tube is sufficient to destroy the exactness 

 of even this observation ; and in the most calm weather, the 

 column is never perfectly at rest, but shows small though sensi- 

 ble oscillations. The approach oif the observer, and the handling 

 of the instrument, communicate to some of its parts a heat 

 which has not time to distribute itself equally to the other parts, 

 and the error resulting is generally something, if not equal to the 

 1-lOOOth of an inch : in fact, attempts at excessive accuracy in 

 the observation have in general only the effect of making a dis- 

 play of figures in the result, which the real capability of the 

 instrument will not warrant. 



We ought surely to take the different means of ameliorating 

 observations in the order of their importance ; and no one who 

 has ever employed the barometer with attention is ignorant that 

 of all the errors imputable to the instrument, the most frequent 

 and the most considerable are those which arise from a false 

 indication of the temperature of the mercury. Now we do not 

 here discuss minute fractions of the smallest divisions ; one 

 degree of the centigrade thermometer corresponds, on the baro- 

 metric scale, with more than 3-lOOths of an inch ; and in eleva- 

 tion, with a difference of more than 3 feet. While we are at one 

 end of the mercurial column attempting to estimate microscopic 

 intervals, we must not forget that at the other, a slight and often 

 inevitable mistake may make the observation lose in exactness 

 ten times what it gains by the accuracy of the vernier. 



The greatest improvement which could be introduced, and 

 which would confer the greatest honour on artists, would consist 

 in finding a method of uniting the correctional thermometer to 

 the column of mercury in so close and immediate a manner, that 

 its indications should, at all times, and under all circumstances, 

 be the exact measure of the temperature of the column. 



The precautions have hitherto been confined to inclosing the 

 bulb of the thermometer in the mounting of the instrument, so 

 that the variations of external temperature may be considered as 

 affecting it only through the intervention of the mounting. This 

 does not dispense with the observer's having an eye upon all that 

 may disturb the accordance of the two instruments. Rapid 

 changes of temperature are above all to be suspected, for the 

 correctional thermometer always marks them before the entire 

 mass of the instrument partakes in them ; and we can do no- 

 thing better than to shelter the barometer as far as we can from 

 the influence of these changes. 



The barometer then must be furnished with a good thermo- 

 meter well joined to its mounting, and we must never fail to 

 combine its indications with those of the barometer. The extent 



